Klasifikasi Data FILM untuk Menentukan Rekomendasi FILM untuk Anak Menggunakan Algoritma C5.0

Films are a significant medium that greatly influences everyday life, both positively and negatively. One of the challenges that arise is the mismatch between films and age restrictions, especially for children. This study aims to apply data mining techniques using the C5.0 Algorithm to classify films based on age restrictions to assist parents in selecting appropriate films for their children. This research uses data from the Kaggle website and follows several stages, including data preprocessing, feature selection, dataset splitting, model building, and analysis. Two experimental scenarios were conducted in this study. The first scenario used attributes selected through chi-square feature selection, resulting in an accuracy of 89.7%. The second scenario used attributes selected by the Pearson correlation feature selection method, achieving an accuracy of 91.7%. Although there was an improvement in accuracy in the second scenario compared to the first, this accuracy was still lower than the accuracy obtained without using feature selection, which reached 91.8%. The C5.0 Algorithm without feature selection provided the highest accuracy, with an accuracy rate of 91.8%. This emphasizes that using feature selection does not always improve classification accuracy and that a larger dataset and more complex data variations can contribute to better model performance.